www.sciencemag.org/cgi/content/full/340/6136/1094/DC1

Supplementary Materials for

Mechanisms of Age-Dependent Response to Winter Temperature in Perennial Flowering of Arabis alpina

Sara Bergonzi, Maria C. Albani, Emiel Ver Loren van Themaat, Karl J. V. Nordström, Renhou Wang, Korbinian Schneeberger, Perry D. Moerland, George Coupland*

*Corresponding author. E-mail: coupland@mpipz.mpg.de

Published 31 May 2013, Science 340, 1094 (2013)

DOI: 10.1126/science.1234116

This PDF file includes:

Materials and Methods Supplementary Text Figs. S1 to S9 Tables S1 to S4 References

Other Supplementary Material for this manuscript includes the following: (available at www.sciencemag.org/cgi/content/full/340/6136/1094/DC1)

Table S1 (Excel)

1

Supporting Online Material

Contents Supplementary Text................................................................................................................2

Microarray hybridization experiments ................................................................................2

Development of an A. thaliana custom array to be used for heterologous hybridizations .2

Optimization of the hybridization conditions for A. alpina studies ..................................3

Experimental design to identify genes differentially expressed in apices of younger plants that do not flower in response to vernalization and older plants that do flower in vernalization ...................................................................................................................4

Microarray data analysis: Pre-processing, probe filtering & differential expression .........5

Material and Methods .............................................................................................................7

Plant material, growth conditions and flowering time measurements...................................7

Analysis of gene transcript levels (quantitative RT-PCR) ....................................................7

Analysis of mature miRNA abundance (quantitative RT-PCR) ...........................................8

In situ hybridization ............................................................................................................9

Identification of SPL genes in A. alpina ..............................................................................9

Sequence alignments and phylogenetic analysis ................................................................ 10

Identification of miR156 precursors in A. alpina ............................................................... 10

A. alpina transformation ................................................................................................... 10

Supplementary Figure Legend .............................................................................................. 11

Fig. S1. PEP2 is a floral repressor and is genetically linked to AaAP2 .............................. 11

Fig. S2. PEP2 mRNA levels in apices of 2-week-old and 8-week-old plants do not differ during exposure to cold although miR172 binding site in PEP2 is conserved .................... 12

Fig. S3. Phylogenetic relationships of A. thaliana and A. alpina SPL genes ...................... 12

Fig. S4. miR156 abundance decreases in aging axillary shoots and low levels correlate with the ability of the shoot to flower in response to cold exposure ........................................... 12

Fig. S5. miR156 downregulation is arrested by cold in A. thaliana and A. alpina delaying flowering potential ............................................................................................................ 13

Fig. S6. Expression of MIR156b under the CaMV35S promoter in pep1-1 background strongly delayed flowering in long days ............................................................................ 13

Fig. S7. Model for flower induction in A. alpina Pajares. .................................................. 14

Fig. S8. Customized A. thaliana Agilent array developed for heterologous hybridizations to A. alpina mRNA ............................................................................................................... 15

Fig. S9. Validation of qPCR technique for miRNA detection ............................................ 16

Supplementary Tables .......................................................................................................... 16

Table S1. Excel file with the analyzed microarray data ..................................................... 16

Table S2. Differentially expressed SPL genes detected by microarray analysis using FDR < 0.05 .................................................................................................................................. 16

2

Table S3. Conservation of synteny between A. thaliana and A. alpina SPL genes. ............ 17

Table S4. Primer list ......................................................................................................... 18

References for Supporting Online Material ........................................................................... 19

Supplementary Text

Here we provide further information about the interspecific microarray experiment presented

in the main text. We exploited the expected sequence similarities between the two species by

hybridizing A. alpina mRNA on a customized A. thaliana array. We describe the design of the

custom array and the tests we performed in order to optimize the heterologous hybridization

conditions. Moreover we give a more extensive description of the experimental design as well

as of the dedicated data analysis.

Microarray hybridization experiments

Development of an A. thaliana custom array to be used for heterologous hybridizations

To develop an Arabidopsis array to be used for heterologous hybridization studies we chose

the Agilent microarray technology that includes 60-mer oligos and therefore offers higher

sensitivity for hybridization of RNAs from closely related species than other technologies that

use shorter oligos, such as Affymetrix. The Agilent 3 array includes approximately 27,000

annotated genes and around 10,000 non-annotated Arabidopsis genomic regions which are

transcribed (http://www.genomics.agilent.com/). Agilent 3 is a standard commercial platform

based on oligos derived from the 3’UTRs. Therefore, a customized array was designed in

which, to increase the probability of a successful heterologous hybridization, every gene is

represented with an oligo-set generating a total of 244,000 oligos. Oligos were designed using

the eArray application based on the Arabidopsis DNA sequence, starting from the less

conserved 3´ end and going in the direction of the 5´ end of every gene. Most of the genes

were represented on the array by either 10 or 7 oligos. The distance between two oligos on

the same gene varied based on the optimization protocol of eArray that accounts for low

3

complexity and repetitive regions. The calculated median distance between the end and the

start of two consecutive oligos was 27 nucleotides. In addition, oligos specific for 21 known

genes of A. alpina were also included on the array as controls.

Optimization of the hybridization conditions for A. alpina studies

To test the suitability of the customized array for cross-species hybridization, A. alpina

genomic DNA and A. thaliana genomic DNA were hybridized to the custom array and the

hybridization intensities compared. This approach allowed an estimation of the number of

genes whose expression could be expected to be detected by hybridizing cDNA. DNA was

isolated using the DNeasyTM Kit (Qiagen). For the majority of the genes on the array, 5 or

more oligos hybridized when A. alpina DNA was used as probe (fig. S8A). Around 250 A.

thaliana genes did not hybridize to A. alpina DNA, so no hybridization was detected to any of

the gene-specific oligos (fig. S8A, red bar). When A. thaliana DNA was used as target most

of the genes hybridized to all the oligos (fig. S8B). In addition, the oligos designed at the 3’

end of the genes were generally those that did not hybridize to A. alpina, whereas for A.

thaliana the position of the oligo in the gene did not influence the hybridization (fig. S8, C

and D). These results were expected because the 3’ ends are generally less conserved among

species. Moreover, this supports the choice of the custom array design in which every gene is

represented by an oligo-set instead of a single oligo.

To test the RNA hybridization method, a cheaper commercially available 22K Agilent array

was used. As for the custom array, the oligos on this array are in situ synthesised and 60-mer

long. To identify the optimal hybridization conditions for the heterologous RNA

hybridization, we tested two different hybridization temperatures. The standard Agilent array

protocol uses 65°C as hybridization temperature for homologous RNA hybridization. Two

lower temperatures were tested for both homologous and heterologous hybridizations using A.

thaliana and A. alpina mRNAs. The same biological sample was used for A. thaliana,

4

hybridized to all the arrays and considered as technical replicate. For A. alpina two different

biological samples were hybridized as duplicates. An overview of the hybridizations is shown

in fig. S8E. The signal intensity for A. alpina was in general lower than the signal intensity for

A. thaliana but the differences were not striking (fig. S8E). The medians of intensities for A.

alpina hybridizations were slightly higher at 50°C compared to 60°C (fig. S8E, left side). In

addition, plotting the two hybridizations in a scatter plot showed that additional oligos

hybridized at 50°C compared to 60°C (fig. S8F). However, this increase in hybridizing oligos

was also observed in A. thaliana DNA hybridizations and is therefore likely to be non-specific

(fig. S8G). Since the same biological sample of A. thaliana was used in duplicate for all the

hybridizations at both temperatures, reproducibility could be tested at different temperatures

(fig. S8, H and I). Reproducibility was higher at 60°C compared to 50°C. Taken together the

test hybridization results showed that temperature slightly influences the intensity of

hybridization of A. alpina targets (fig. S8E). Sensitivity increases when hybridizing at 50°C

compared to 60°C, but this likely reflects at least in part non-specific hybridization since this

can also be observed in A. thaliana (fig. S8, F and G). Therefore, the subsequent experiments

were performed at 60°C in order to gain in reproducibility and possibly in specificity.

Hybridization times, washing condition, and scanning of the arrays were performed according

to the Agilent protocols.

Experimental design to identify genes differentially expressed in apices of younger

plants that do not flower in response to vernalization and older plants that do flower in

vernalization

A common reference design was used in the microarray experiment in which a pool of the

RNAs of all 16 samples was used as the common reference (1). Every slide was therefore

hybridized with the common reference sample, labelled with the green fluorescent dye

Cyanine 3 (Cy3), and with one of the biological samples, labelled with the red fluorescent dye

5

Cyanine 5 (Cy5). The apices collected at different stages of development and in different

conditions were pooled and homogenized in liquid nitrogen. Pooled apices from plants grown

for 2 weeks in long day formed the 2-week-old sample before exposure to cold (2wo). Pooled

apices from plants grown for 8 weeks in long days formed the 8-week-old sample before

exposure to cold (8wo). Pooled apices from plants grown for 2 weeks in long days and then

shifted to 4°C for 4 weeks formed the 2-week-old sample during cold treatment (2wo + cold).

Pooled apices from plants grown for 8 weeks in long days and then shifted to 4°C for 4 weeks

formed the 8-week-old sample during cold treatment (8wo + cold). Each pool was considered

as a single biological replicate. Four biological replicates were collected for each condition.

Total RNA was isolated using the RNeasyTM Mini Kit (Qiagen). At this stage, double column

purification was performed using the same kit and following the manufacturer’s instructions.

Approximately 100 mg of tissue was used for the RNA extraction of every biological

replicate. Genomic DNA was afterwards digested using the DNA-freeTM Kit (Ambion). RNA

quality and integrity were measured with the Bioanalyzer (Agilent). The Agilent Quick Amp

Labeling kit was used to synthesize Cy3 and Cy5 labeled cRNA. For each RNA sample, 500

ng were used for the labelling. Afterwards the concentration and incorporation of the cRNA

and dyes were measured with the Nanodrop. For the hybridization, 2000 ng of Cy3 and Cy5

labeled cRNA were used for further fragmentation and hybridization to the custom microarray

(labeling and hybridizations were performed at Service XS, Leiden, the Netherlands).

Microarray data analysis: Pre-processing, probe filtering & differential expression

To obtain a representative gene expression value for each gene on each array, we corrected

for the background signal, normalized the log-ratio and intensity distributions, filtered low

intensity oligos and summarized the oligos into one single value per oligo-set. To normalize

the data we first normalized the log-transformed ratio distribution within each array using

intensity-dependent normalization (function “normalizeWithinArrays”, method “loess”, R

6

package “limma”). To make the values comparable between arrays we applied quantile

normalization to ensure that the common reference channel has the same intensity distribution

across arrays (function “normalizeBetweenArrays”, method “Gquantile”, R package

“limma”). We then excluded oligos with maximum intensity lower than 7.3 in the common

reference channel of all 16 arrays. This threshold was based on the 75th percentile of the

intensity distribution observed for intron coding oligos, which served as a control for the

interspecific hybridization. In total, 27607 (91%) genes have at least one 60-mer exon coding

oligos with a maximum expression in the common reference channel above 7.3. The

summarization step for genes with multiple oligos was performed using the Robust Multiarray

Average (RMA) method (2) on the normalized log-transformed red channel intensities. In

brief, RMA uses median polish to find a robust median expression value for each gene on

each channel. After the summarization step log-ratios were calculated and used in

downstream analyses.

A 2x2 factorial design was used with vernalization and age as factors. For each gene a linear

model was fitted with a factor containing the 4 conditions as explanatory variable. To find

genes differentially expressed between two conditions we employed a moderated t-test (3)

and adjusted the resulting p-values to correct for multiple hypothesis testing using the

Benjamini-Hochberg false discovery rate. The heatmap was generated with Pearson

correlation as distance measure, complete linkage as agglomeration method; rows were scaled

according to the function ‘heatmap.2’ from the R-package ‘gplots’.

7

Material and Methods

Plant material, growth conditions and flowering time measurements

The A. alpina accession Pajares (referred to in the text as wild-type) and pep1-1 mutant were

previously described (4). The pep2-1 allele was initially identified using a next generation

sequencing approach and described as floral defective 1 (fde1), due to its floral phenotype (5).

The additional pep2 alleles (pep2-2 to pep2-7) were obtained from the enhancer screen of the

pep1-1 mutant according to the protocol described in (4).

For all experiments seeds were first cold–imbibed for 2-4 days at 4°C. Plants were routinely

grown in control conditions in long days (16 h light/8 h dark) unless differently indicated.

Light was provided by fluorescent tubes complemented by incandescent bulbs to increase the

proportion of far red light. The temperatures ranged from 20°C during the day to 18°C during

the night, unless differently indicated. Cold and vernalization treatments were performed at

4°C and in short-day conditions (8 h light/16 h dark) to mimic the natural winter conditions.

For A. thaliana experiments Columbia plants were grown in short days at 20°C and exposed

to cold as A. alpina plants.

Flowering time was measured both as Days to Flower and as Total Leaf Number (TLN) from

the time seeds were sown on soil. In both cases at least 12 individuals were scored. Data

presented as means ± standard deviation.

Expression profiling by microarray and qRT-PCR was performed using RNA extracted from

main shoot apices of 10 to 30 individuals at ZT 8 (zeitgeber time).

Analysis of gene transcript levels (quantitative RT-PCR)

Total RNA was extracted from apices using the miRNeasyTM Mini Kit (Qiagen) followed by

DNA digestion using the DNA-freeTM Kit (Ambion). miRNeasyTM Mini Kit allows

purification of total RNA including RNA from approximately 18 nt in length. The same RNA

8

was then used for mRNA, pre-miRNA and miRNA quantifications (see next section). For

mRNA, cDNA synthesis was performed using the OligodT primer and the Superscript II

reverse transcriptase enzyme (Invitrogen). To detect expression of pri-miR156a, cDNA was

synthesized with OligodT and a specific reverse oligos designed on the stem loop (stem loop

oligo - SLO) at concentration of 0, 5 µM. In both cases cDNAs were diluted to 150 µl with

water and 3 µl were used as a template for qRT-PCR using a BioRad iQ5 apparatus and

SYBR-green detection. The A. alpina homologue of PROTEIN PHOSPHATASE 2A

(AaPP2A) was used as reference gene to normalize the varying amounts of cDNA between

samples. Two or more biological replicates were performed for every experiment and when

consistent one replicate was shown. Error bars represent standard error for technical

replicates. Primer sequences are listed in Table S4.

Analysis of mature miRNA abundance (quantitative RT-PCR)

Starting from the total RNA we performed 1st strand synthesis of the mature form of miRNAs

as previously described in (6) and modified by (7). Additionally, this method was used to

detect miR156 and miR172 in apices of Arabidopsis plants growing in short-day conditions

and could reproduce the results previously reported using Northern blot analysis (fig. S9) (8).

In short, the mature forms of miR156 and miR172 were reverse transcribed starting from 200

ng of total RNA and using two different key-primers each composed of a universal sequence

of 35 amino acids at the 5´ end and by 8 nucleotides complementary to the miR156 or the

miR172 at the 3´ end. The reverse primer for the small nucleolar RNA 101 (snoR101), which

is the reference used to normalize the qRT-PCR, was also included in the reverse

transcription. The product from the 1st strand synthesis was diluted to 200 µl. In a second step,

3 µl of the reverse transcription products were quantified by SYBR-green based qRT-PCR

using a miRNA-specific forward primer and a generic universal reverse primer designed on

the key-primer. Since the mature forms of miR156 and miR172 are likely highly conserved

9

between A. alpina and A. thaliana, the same primer sets were used for the detection of the two

miRNAs in both species. Different primer pairs were used for the A. thaliana and A. alpina

snoR101s. Two or more biological replicates were performed for every experiment (with the

exception of the extended cold treatment experiments) and when consistent one replicate was

shown. Error bars represent standard error for technical replicates. Primer sequences are listed

in Table S4.

In situ hybridization

Longitudinal sections of plant apices were probed with digoxigenin (DIG)-labelled antisense

mRNA probes. The probe to detect PEP2 transcript was PCR- amplified from cDNA using

specific primer pairs that amplify part of the cDNA, with T7 RNA polymerase binding site

attached to the reverse primer (5´- TAATACGACTCACTATAGGG -3´) and T3 RNA

polymerase binding site attached to the forward primer (5´- ATTAACCCTCACTAAAGGGA

-3´). RNA in situ hybridization was carried out as described by (9) with small modifications

described in (10).

For detection of miR156 and miR172 expression patterns we purchased the miR156g and

miR172b antisense LNA (locked nucleic acid, Exiqon, Vedbaek, Denmark) oligonucleotides

that were 5´-DIG and 3´-DIG labeled. LNA-based miRNA in situ hybridization was carried

out largely according to the same procedure as for mRNA in situ hybridization. We used 10

pmol of miR172b LNA probe or 2.5 pmol of miR156g LNA probe for every section.

Hybridization was performed at 50°C. Primers and LNA probe sequences are listed in Table

S4.

Identification of SPL genes in A. alpina

The Arabidopsis SPL gene list was obtained from the DATF SBP-box gene family database

(11). We made use of the data from the on-going A. alpina sequencing project (unpublished),

which includes sequencing of both genome and transcriptome, and performed HMMER

10

search against the A. alpina predicted proteome using the extracted PFAM SBP domain

(http://hmmer.janelia.org/). All hits with E-values below 1e-30 were extracted.

Sequence alignments and phylogenetic analysis

Following the identification of the 15 SPL genes, multiple alignments of the amino acid

sequences were generated using the Neighbor-Joining method (12). The bootstrap consensus

tree has been inferred from 10000 replicates (13). Phylogenetic analyses were conducted in

MEGA5 (14).

Identification of miR156 precursors in A. alpina

To identify the A. alpina precursors of miR156 we follow an approach previously described

by (15). The sequences of the mature miR156s from A. thaliana and several other species

obtained from the miRBase database (www.miRBase.org) were used to BLAST against the

genome of A. alpina (Alpina alpina whole-genome consortium, unpublished). In total, 3,029

hits were recovered using a cut off of 90% similarity. Based on synteny conservation the

putative orthologous genes to the A. thaliana MIR156 genes were identified among the

obtained hits. The sequences were extended on both sides of the miR156 binding site in order

to obtain the full length of the A. alpina pre-miR156s. The presence of the reverse

complement miR156 sequence was assessed for each candidate gene. Estimations of the

lengths of the A. alpina precursors were based on the alignment with the A. thaliana pre-

miRNAs.

A. alpina transformation

Plasmids containing 35S:MIR156b and 35S:MIM156 (16, 17) were obtained from Detlef

Weigel (MPI for Developmental Biology, Tübingen). 35S:MIR156b is in the pMLBART (18)

and was introduced into the Agrobacterium tumefaciens strain GV3101 (pMKRK).

35S:MIM156 construct is in the pGREEN vector and was introduced into the A. tumefaciens

11

strain GV3101 (pSOUP). Plasmids were introduced into A. alpina by floral dip method (19).

T1 transformants were identified on the basis of BASTA resistance.

Supplementary Figure Legend

Fig. S1. PEP2 is a floral repressor and is genetically linked to AaAP2

(A) The early-flowering phenotype of the pep2-1 mutant co-segregated with AaAP2. The

pep2-1 mutant was backcrossed to the accession Pajares and the F1 progeny were self-

fertilized. Flowering was tested in 93 F2 plants 100 days after germination and growth in long

days. Plants were scored as flowering or non-flowering and segregated with a ratio not

significantly different to 1:3 (17 flowering : 76 non-flowering). The 17 flowering plants were

all homozygous for the mutation in AaAP2 detected in (5) and showed floral homeotic

defects. However, the flowering phenotype was leaky and 3 plants homozygous for the

mutation in AaAP2 did not flower after 100 days in long-days. Gray color: non flowering

plants; white color: plants flowering in long days.

(B) Conservation of synteny of A. alpina genomic region carrying APETALA 2 (AaAP2) with

A. thaliana. GATA plot of A. thaliana genomic region at chromosome 4 flanking At4g36920

(AP2) with the A. alpina region containing AaAP2.

(C) pep2-1 floral phenotype as in (5)

(D) PEP2 can repress flowering independently of PEP1. Flowering time measured as total

leaf number (TLN) of pep1 and pep2-2 pep1-1.

(E and F) pep2-1 enhances the early-flowering phenotype of pep1-1. Flowering time of pep1-

1 pep2-1 double mutants compared to pep1-1 and pep2-1 single mutants measured as days to

flower (E) and total leaf number (TLN) (F).

12

Fig. S2. PEP2 mRNA levels in apices of 2-week-old and 8-week-old plants do not differ

during exposure to cold although miR172 binding site in PEP2 is conserved

(A) PEP2 mRNA in apices of plants exposed to cold (4°C). Plants grown for 2 (white) or 8

(black) weeks prior to cold.

(B) Alignment of the miR172 binding site AP2 and AP2-like genes in A. thaliana and

homologous region in PEP2.

Fig. S3. Phylogenetic relationships of A. thaliana and A. alpina SPL genes

(A) Unrooted tree of 17 A. thaliana and 16 predicted A. alpina SPL genes using predicted

amino acid sequences (MEGA5 software). Neighbour-joining algorithm was used with a

bootstrapping analysis of 10,000 reiterations. The percentage of replicate trees in which the

associated taxa clustered together in the bootstrap test is shown next to the branches when

higher than 50%. Asterisks: genes encoding the miR156 binding site.

(B) Sequence comparison of the miR156 binding site between A. thaliana and A. alpina SPL

genes. Red: nucleotide changes.

Fig. S4. miR156 abundance decreases in aging axillary shoots and low levels correlate

with the ability of the shoot to flower in response to cold exposure

(A) miR156 abundance in the apices of axillary shoots formed at the 1st and 2nd node of a 5-

week-old plant. miR156 levels were followed for 5 weeks between 5 weeks and 10 weeks

after germination. The miR156 level in the main shoot apex has already reached trough levels

by 5 weeks after germination (Fig. 3B), indicating that the reduction in these axillary shoots

occurs independently of the main shoot apex.

(B) 5-week-old wild-type plant. Axillary shoots are visible at the axes of the 1st and 2nd leaf.

(C) 8-week-old wild-type plant.

13

(D) 5-week-old plant exposed to cold temperature for 12 weeks and returned to long days for

5 weeks. Axillary shoots at 1st and 2nd nodes remained vegetative.

(E) 8-week-old plant exposed to cold temperature for 12 weeks and returned to long days for

5 weeks. The axillary shoots formed at the lower nodes flower.

Fig. S5. miR156 downregulation is arrested by cold in A. thaliana and A. alpina delaying

flowering potential

(A) Extended cold treatment does not induce flowering of young seedlings. An A. alpina plant

grown for 2 weeks in long days and then exposed to cold for 60 weeks. Scale bar = 1 cm.

(B) A. alpina plant which acquired sensitivity to flower upon vernalization after growing for 5

weeks in long days. Scale bar = 1 cm. Although able to flower this plant has produced fewer

leaves than the vegetative plant shown in (A) grown for 60 weeks in cold.

(C) miR156 abundance in A. thaliana Columbia plants grown either continuously at 20°C in

short days (black) or that have been transferred after 2 weeks at 20° C to cold for 33 weeks

(white). A. thaliana plants started bolting after 33 weeks at 4°C.

(D) Expression of the A. alpina pre-miR156a precursor in apices of 3-week-old (white) and 8-

week-old (black) plants before and at the end of cold treatment.

Fig. S6. Expression of MIR156b under the CaMV35S promoter in pep1-1 background

strongly delayed flowering in long days

(A) miR156 abundance in main apices of pep1-1 plants (black) growing in long days (LDs)

and comparison with the abundance in wild-type plants (gray).

(B) 35S:MIR156b pep1-1 plants grown for 7 months in long days. Only a few axillary shoots

initiated flowering after several months. Scale bar = 7 cm.

14

Fig. S7. Model for flower induction in A. alpina Pajares.

A. alpina Pajares flowers only when exposed to cold at the age of 5 weeks or older. An age-

dependent timer based on miR156 abundance ensures SPL transcripts, which are essential for

flowering to occur, to increase only in older plants. In parallel, PEP2 represses flowering until

plants are exposed to cold both by increasing PEP1 expression and by a PEP1 independent

mechanism. Cold temperature plays a second role by strongly delaying reduction of miR156

levels in aging plants, therefore preventing flowering under winter conditions. Based on a

different experimental design miR156 was previously reported to increase under lower

ambient temperatures (20). We propose that low temperatures rather delay the rate of

reduction of miR156 precursor transcription as the plant ages.

Panel A: young plants exposed to warm temperatures do not flower because both PEP2/PEP1

and miR156 pathways repress flowering; panel B: after exposure to a prolonged cold period

PEP1 mRNA levels are downregulated but young plants still do not flower because cold also

slows down the decrease of miR156 levels and therefore flowering is repressed; panel C:

during growth in long days miR156 levels are gradually decreased as plants get older, but

plants do not flower because the PEP2/PEP1 pathway is still active repressing flowering;

panel D: in older plants miR156 levels are low and AaSPL genes (AaSPLs) are expressed.

When these plants are vernalized the reduction of PEP1 mRNA allows flowering to occur

because both repressive pathways are switched off. Under these conditions miR172

abundance also increases in the inflorescence meristem and the inner whorls of the floral

primordium correlating with flower development. PEP2 mRNA expression falls in the

inflorescence meristem and rises in the outer whorls of the floral primordium. x: unknown

proposed age-related factor that increases with age leading to the repression of MIR156

transcription. Black: expressed gene; grey: not expressed.

15

Previously AaTFL1 was also shown to repress flowering when young A. alpina plants are

exposed to vernalization (21). AaTFL1 is not included in the figure, which illustrates the work

done in the present paper. However, as TFL1 acts in A. thaliana to negatively regulate the

floral identity genes APETALA1 and LEAFY, which are positively regulated targets of SPL

transcription factors (8, 22), we propose that AaTFL1 sets a threshold for AaSPL activation of

flowering. Therefore in the absence of AaTFL1 lower levels of AaSPL that are achieved

earlier in development promote flowering.

Fig. S8. Customized A. thaliana Agilent array developed for heterologous hybridizations

to A. alpina mRNA

(A to D) DNA hybridizations on A. thaliana customized Agilent array. (A and C) A. alpina

DNA hybridization. (B and D) A. thaliana DNA hybridization. (A and B) Barplots

representing the number of genes with 1 oligo of the oligo-set, 2 oligos, 3 oligos etc.

hybridizing with intensity higher than 6 on log2-scale (the background signal intensity on

log2-scale was equal to 6). Red bar: A. alpina genes that did not hybridize with any oligo. (C

and D) Boxplots representing the hybridization intensities according to the position of the

oligos along the gene. x-axis: position of the oligos from the most 3` end (number 1) to the

most 5´ end (number 10). y-axis: signal intensity in log2.

(E to I) RNA hybridizations on A. thaliana customized Agilent array. (E) Boxplot of A.

alpina and A. thaliana hybridization at the two temperatures tested, 50°C and 60°C. x-axis:

arrays. y-axis: signal intensities in log2. (F) Scatter plot representing the A. alpina RNA

hybridization intensities at 50°C (x-axis) and 60°C (y-axis). Red line: Loess curve of the

observed intensities; blue line: diagonal representing perfect correlation. (G) Scatter plot

representing the A. thaliana RNA hybridization intensities, the rest as in (F). (H and I)

Scatter plots for the two technical replicates of A. thaliana hybridization at 50°C (H) and at

60°C (I).

16

Fig. S9. Validation of qPCR technique for miRNA detection

miR156 (black) and miR172 (white) expression measured by qPCR in apices of A. thaliana

Columbia plants growing in short days (SDs) until flowering.

Supplementary Tables

Table S1. Excel file with the analyzed microarray data

Table S2. Differentially expressed SPL genes detected by microarray analysis using FDR

< 0.05

Differentially expressed genes of the SPL family are shown in two comparisons:

(A) Apices of 8-week-old plants (8wo) vs 2-week-old plants (2wo) growing in long days

before experiencing cold.

(B) Apices of 8-week-old plants (8wo) vs 2-week-old plants (2wo) exposed to cold for 4

weeks (+ cold).

FDR < 0.05; Ranking according to statistical significance (FDR).

A. Comparison between apices of 8wo vs 2wo plants before cold treatment

ranking position SPL gene Fold Changes FDR (<0.05)

11 SPL10 1.59 4.17E-07 17 SPL9 2.10 6.73E-07

169 SPL6 1.44 3.03E-05 247 SPL13 1.93 6.42E-05 415 SPL11 1.38 2.14E-04 454 SPL2 1.56 2.84E-04 575 SPL3 1.92 5.12E-04

17

B. Comparison between apices of 8wo vs 2wo plants upon cold treatment

ranking position SPL gene Fold Changes FDR (<0.05) 2 SPL10 1.96 4.04E-08 4 SPL9 2.30 4.96E-07

59 SPL6 1.54 1.52E-05 144 SPL13 2.02 6.52E-05 219 SPL11 1.43 1.62E-04 331 SPL5 2.23 3.24E-04 420 SPL2 1.54 5.03E-04 1455 SPL15 1.30 6.56E-03

Table S3. Conservation of synteny between A. thaliana and A. alpina SPL genes.

For every AtSPL gene, the table lists the A. thaliana flanking genes for which we can find a

conserved homologue in the flanks of the corresponding AaSPL gene. Insertions are not

shown.

AaSPL AtSPL

AaSPL1 At2g47050 At2g47060 At2g47070 At2g47090 At2g47100 AaSPL2 At5g43270 AaSPL3 At2g33793 At2g33800 At2g33810 At2g33820 At2g33835 AaSPL4 At1g53130 At1g53140 At1g53160 At1g53163 At1g53165 AaSPL5 At3g15251 At3g15260 At3g15270 At3g15280 At3g15290 AaSPL6 At1g69130 At1g69160 At1g69170 At1g69180 At1g69200 AaSPL7 At5g18810 At5g18820 At5g18830 At5g18840 At5g18850 AaSPL8 At1g02050 At1g02060 At1g02065 At1g02070 At1g02080 AaSPL9 At2g42180 At2g42190 At2g42200 At2g42210 At2g42220

AaSPL10/ 11 At1g27320 At1g27340 At1g27360/At1g27370 At1g27380 At1g27385 AaSPL12 At3g60010 At3g60020 At3g60030 At3g60040 At3g60050

AaSPL13A/B At5g50540 At5g50560 At5g50570/ At5g50560 At5g50580 At5g50590 AaSPL14 At1g20980 At1g20990 AaSPL15 At3g57890 At3g57910 At3g57920 At3g57930 At3g57940 AaSPL16 At1g76560 At1g76570 At1g76580 At1g76590 At1g76600

18

Table S4. Primer list

ID Gene Sequence (5´-3´) Purpose Sense E054 PEP2 TTG AAG CTG CTA GGG CTT ATG qRT-PCR F E055 PEP2 TTG GAT TCC CTG ATG ACT CG qRT-PCR R E005 PEP2 TGA GGA GAA TCA AAC CAG AGG PCR cDNA F E026 PEP2 TTT TTG GCC TAA TTA ACA AGA GG PCR cDNA R B279 PEP2 (T3)TGTGGGATCTAAATGATGCAC in situ F B280 PEP2 (T7)CGGTAGCAAGATCCGAAGAC in situ R W132 PEP1 CTTGTCGTCTCCTCCTCTGG qRT-PCR F W133 PEP1 ACTACGGCGAGAGCAGTTTC qRT-PCR R W268 AaSOC1 GCTTTCAGTGCTTTGTGATGC qRT-PCR F W269 AaSOC1 GGATGCTTCGAGTTGTTCGAT qRT-PCR R W578 AaLFY ACG CCG TCA TTT GCT ACT CT qRT-PCR F W579 AaLFY TTT GCG TCA TCG TCT GTC TC qRT-PCR R B217 AaPP2A AGTATCGCTTCTCGCTCCAG qRT-PCR F B218 AaPP2A AACCGGTTGGTCGACTATTG qRT-PCR R F408 miR172 GTTGGCTCTGGTGCAGGGTCCGAGGTATTCGCACCAGAGCCAACATGCAG Key primer universal R B166 miR172 AGAATCTTGATGATGC qRT-PCR F F409 miR172 GGCGGAGAATCTTGATGATG qRT-PCR R B163 miR156 CGCGAGCTCAGAATTAATACGACTCACTATACGCGGTGCTCAC Key primer universal R B164 miR156 TGACAGAAGAGAGTG qRT-PCR F B165 miR156 CGCGAGCTCAGAATTAATACGA qRT-PCR R B148 Aa snoR101 TTTTACAGGTAAGTTCTTGTTG qRT-PCR F B149 Aa snoR101 AGCATCAGCAAACCAATAGTT qRT-PCR R B146 At snoR101 CTTCACAGGTAAGTTCGCTTG qRT-PCR F B147 At snoR101 AGCATCAGCAGACCAGTAGTT qRT-PCR R B281 pre-miR156a TGGGACAAAAGAAACGCAAAG qRT-PCR F B282 pre-miR156a GACAGGCCAAAGAGATCAGC qRT-PCR (SLO) R

30126-15 miR156g GTGCTCACTCTCTTCTGTCG LNA probe - 30063-15 miR172b GTGAATCTTAATGGTGCTGC LNA probe -

B156 AaSPL9 CAGATAAACGGTGGAACGAC qRT-PCR F B157 AaSPL9 CTTTGATGCAAAGAGAAGTAG qRT-PCR R B258 AaSPL10 ACACAATCCCCTGATCGAAG qRT-PCR F B259 AaSPL10 AAACCGAGGATAACCGAGTAAC qRT-PCR R B260 AaSPL6 GATTTATTGGATTGGTGAGATGG qRT-PCR F B261 AaSPL6 TCTGAAACCCTGAGAAAGAAGC qRT-PCR R B185 AaSPL13 AGAGCTCGAGAACAGCATCG qRT-PCR F B186 AaSPL13 CTGCACTCGCATTCTCAAAC qRT-PCR R B246 AaSPL2 CGGTGTAGGCCAGTTTAATG qRT-PCR F B247 AaSPL2 TGGGCAATAAAGAAGAAAGGAC qRT-PCR R B248 AaSPL15 TGCACCAAGAGCTTTACCTG qRT-PCR F B249 AaSPL15 GGCTTAAAGATCAAAAGCCAAAG qRT-PCR R B268 AaSPL3 TGCTTGCTCTCTTCTGTCAGTC qRT-PCR F B269 AaSPL3 TTCCAACAAGCATTTACTTAGTCTC qRT-PCR R B244 AaSPL8 AGTCCTCCTCCCCAAACG qRT-PCR F B245 AaSPL8 GCCTAGCCACTGGAGAAGAAG qRT-PCR R B254 AaSPL7 GGTCTGTGCGGTTCTCTACC qRT-PCR F B255 AaSPL7 AGCTCCCTTCAACCTTTTTG qRT-PCR R B160 AaSPL5 CCAGCTGGCTCTCTCTCTTC qRT-PCR F B161 AaSPL5 TGCGGATAGACATCCATTATAGG qRT-PCR R B266 AaSPL4 TAATCGAAAGCCACGGTCAC qRT-PCR F B267 AaSPL4 CCCAATCGATTTTATTCATATCTTG qRT-PCR R

19

References for Supporting Online Material

1. Y. H. Yang, T. Speed, Design issues for cDNA microarray experiments. Nat Rev Genet 3, 579

(2002). 2. R. A. Irizarry et al., Summaries of Affymetrix GeneChip probe level data. Nucleic Acids Res 31,

e15 (2003). 3. G. K. Smyth, Linear models and empirical bayes methods for assessing differential expression

in microarray experiments. Stat Appl Genet Mol Biol 3, Article3 (2004). 4. R. H. Wang et al., PEP1 regulates perennial flowering in Arabis alpina. Nature 459, 423

(2009). 5. K. Nordström et al., Mutation identification by direct comparison of whole-genome sequencing data from mutant and wild-type individuals using k-mers. Nat. Biotechnol. 10.1038/nbt.2515 (2013). 6. H. Yang et al., A novel real-time polymerase chain reaction method for high throughput

quantification of small regulatory RNAs. Plant Biotechnol J 7, 621 (2009). 7. S. Xing et al., miR156-targeted and nontargeted SBP-box transcription factors act in concert

to secure male fertility in Arabidopsis. Plant Cell 22, 3935 (2010). 8. J. W. Wang et al., miR156-Regulated SPL Transcription Factors Define an Endogenous

Flowering Pathway in Arabidopsis thaliana. Cell 138, 738 (2009). 9. D. P. Jackson, In situ hybridization in plants. M. P. P. A. P. Approach, Ed., (D.J. Bowles, S.J.

Gurr and M.McPhereson, Eds, Oxford, ed. Oxford University Press, 1992). 10. I. Searle et al., The transcription factor FLC confers a flowering response to vernalization by

repressing meristem competence and systemic signaling in Arabidopsis. Genes Dev 20, 898 (2006).

11. A. Guo et al., DATF: a database of Arabidopsis transcription factors. Bioinformatics 21, 2568 (2005).

12. N. Saitou, M. Nei, The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol Biol Evol 4, 406 (1987).

13. J. Felsenstein, Confidence-Limits on Phylogenies - an Approach Using the Bootstrap. . Evolution 39, 783 (1985).

14. K. Tamura et al., MEGA5: molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance, and maximum parsimony methods. Mol Biol Evol 28, 2731 (2011).

15. E. Bonnet et al., Detection of 91 potential conserved plant microRNAs in Arabidopsis thaliana and Oryza sativa identifies important target genes. Proc Natl Acad Sci U S A 101, 11511 (2004).

16. J. M. Franco-Zorrilla et al., Target mimicry provides a new mechanism for regulation of microRNA activity. Nat Genet 39, 1033 (2007).

17. R. Schwab et al., Specific effects of microRNAs on the plant transcriptome. Dev Cell 8, 517 (2005).

18. A. P. Gleave, A versatile binary vector system with a T-DNA organisational structure conducive to efficient integration of cloned DNA into the plant genome. Plant Mol Biol 20, 1203 (1992).

19. S. J. Clough, A. F. Bent, Floral dip: a simplified method for Agrobacterium-mediated transformation of Arabidopsis thaliana. Plant J 16, 735 (1998).

20. H. Lee et al., Genetic framework for flowering-time regulation by ambient temperature-responsive miRNAs in Arabidopsis. Nucleic Acids Res 38, 3081 (2010).

21. R. Wang et al., Aa TFL1 confers an age-dependent response to vernalization in perennial Arabis alpina. Plant Cell 23, 1307 (2011).

22. A. Yamaguchi et al., The MicroRNA-Regulated SBP-Box Transcription Factor SPL3 Is a Direct Upstream Activator of LEAFY, FRUITFULL, and APETALA1. Dev Cell 17, 268 (2009).